EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Generative AI in L2 Writing, Feedback, and Assessment: A Systematic Critical Review of Validity, Voice, Equity, and Pedagogical Design

Charly Jerome J

Indian Journal of Language and Linguistics · 28.06.2026 · 7 (2) · 86-120

DOI: 10.54392/ijll2627 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Indian Journal of Language and Linguistics, online erschienen am 28.06.2026. Verfasst von Charly Jerome J. Die Arbeit nennt 119 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Writing in a second language has never been a simple skill to teach or learn, and the arrival of generative artificial intelligence has made the challenge considerably more complicated. Large language models entered classrooms, feedback systems, and assessment platforms so quickly that research could not keep pace with practice. Teachers began revising integrity policies before anyone had studied whether AI use was helping students grow as writers. Assessment bodies began questioning whether their rubrics still measured what they were designed to measure. This systematic critical review brings together 74 empirical studies, systematic reviews, and conceptual papers published between January 2022 and April 2026, a period shaped decisively by the release of ChatGPT and related models. Searches were conducted across five major bibliographic databases using the PRISMA 2020 protocol. The review synthesises evidence on how generative AI tools support or disrupt the writing process, the feedback cycle, and the scoring of written texts in second and additional language contexts. Four tensions run through the entire body of scholarship: questions about what AI scoring systems actually measure, concerns about the erosion of writer voice and identity, unequal access across language backgrounds and resource settings, and the absence of principled frameworks that could help teachers use these tools thoughtfully. The review also documents recurring methodological problems in the current literature, including small samples, short observation windows, and thin theoretical grounding. A research agenda is proposed around construct validity in AI scoring, voice and authorship in human and AI collaborative writing, equity and multilingual access, and ethics in automated assessment at scale. The review offers teachers, institutions, and journal editors a conceptual map of the field, an honest critique of its current weaknesses, and concrete directions for the next phase of scholarship.

Zitieren

APA

J, C. J. (2026). Generative AI in L2 Writing, Feedback, and Assessment: A Systematic Critical Review of Validity, Voice, Equity, and Pedagogical Design. Indian Journal of Language and Linguistics, 7(2), 86-120. https://doi.org/10.54392/ijll2627

Vancouver

J CJ. Generative AI in L2 Writing, Feedback, and Assessment: A Systematic Critical Review of Validity, Voice, Equity, and Pedagogical Design. Indian Journal of Language and Linguistics. 2026;7(2):86-120. doi:10.54392/ijll2627

BibTeX
@article{J2026,
  author = {J, Charly Jerome},
  title = {Generative AI in L2 Writing, Feedback, and Assessment: A Systematic Critical Review of Validity, Voice, Equity, and Pedagogical Design},
  journal = {Indian Journal of Language and Linguistics},
  year = {2026},
  volume = {7},
  number = {2},
  pages = {86--120},
  doi = {10.54392/ijll2627}
}
Kostenlos starten →